Audio OER are not as well-known as video, but they're very useful for:
Audio OER can be delivered as streaming media for the user's convenience, but the sound file (usually an mp3) should also be downloadable so that it can be shared and remixed.
Creating basic audio files will require a microphone, probably a better one than the one that came built into your laptop. The one that is built into ITS's standard issue webcam is fine for most non-musical purposes.
If you are recording music or other sounds that require higher fidelity, talk to a sound technician. If you can't find one among your colleagues, you can ask at a local guitar store or community theater group. Expect to have to pay, and be very clear about your goals, timeline, and budget.
You'll also need audio editing software:
Audio OER must be produced at the best possible quality, to avoid extraneous and irritating noises and to maximize clarity. If possible, use a higher quality microphone than comes with your device or computer, and adjust the microphone's physical distance from the sound source, and its settings, until the sound is clean, clear, and appropriate volume. Instead of "cleaning up the sound" in a sound editor after the fact, record clean in the first place. This is crucial for accessibility for learners who are deaf, hard of hearing, or have auditory processing disorders. It is also less distracting for everyone else.
Do not play music or background noise over people speaking. It's unnecessary, and makes it harder to understand the speaker.
Speak more slowly than usual, and pronounce every syllable clearly. Hurrying and mumbling are problems; an accent or a speech impediment is not. Here is How to speak into a microphone properly - avoiding heavy breathing, swallowing, and popping noises.
Always provide a transcript of verbal content for deaf and hard of hearing users. Many of your hearing users will use it as well!
Depending on whether they will find it useful, also consider using visual renditions of the waveforms of sounds. You can screenshot these in a sound editing application.
Make sure that you record music and non-verbal sound audio OER loud enough for your users to turn the volume way up. Hard of hearing users need to listen at very high volume, and deaf users sometimes know how to perceive intense sound vibrations through touch.
Whatever you create is fine to use, as is any material that you find under a compatible Creative Commons license.
However you may need to get permission to use recordings that you made of other people speaking or performing. This includes not only copyright permission, but also, in some cases, privacy releases.